In the side view monitor system which is in practical use at present, a side camera (such as a CCD camera) is set inside a side mirror, and a real camera image from the side camera is displayed on a monitor screen of a front display unit which is also used for a navigation system.
That is to say, displaying the front lateral area of a vehicle, which is the blind spot of a driver, on the monitor screen allows the driver to recognize the condition of the blind spot.
However, because the side camera is disposed inside the side mirror, there is a high parallax (in a range of 1 m to 2 m) between the camera's point of view and the driver's point of view, and thus the shape of an obstacle or other object in the view from the camera, and that in the view from the driver's seat are completely different.
On the other hand, in a normal case, an experienced driver reconstructs the image from the camera in his/her head, then reconstructs and recognizes a positional relationship with objects, and thus keeps the consistency between the image on the screen and the image viewed by the driver him/her self. On the other hand, an inexperienced driver or a driver who needs to respond instantly may not keep the consistency between the image on the screen and the image viewed from the driver's seat, and thus feels a sense of inconsistency.
In order to eliminate such a sense of inconsistency, a signal of a camera image captured by a blind-spot camera which is provided outside the vehicle body is converted to a virtual camera image which is as if viewed from a position of the driver's point of view so that a converted outside image is generated. In addition, an image of visual recognition area is generated that is a camera image excluding the blind-spot area, the camera image being captured by a driver's point-of-view camera provided near the position of the driver's point of view. A blind-spot image display system for vehicle has been proposed, in which a composite image is obtained by superimposing the converted outside image on the blind-spot area excluded from the image of visual recognition area (for example, see Patent Document 1).